[ 195 ] 



XXIV. Some Remarks on Dr. Siemens's Paper " On Standards 

 of Electrical Resistance, and on the Influence of Temperature on 

 the Resistance of Metals." By A. Matthiessen, F.R.S.* 



l.TN the above paper -f, page 92 (2nd paper), M. Siemens states, 

 J- " It may be asserted without all doubt, that the most expe- 

 rienced and skilful physicists, even with the best instruments and 

 most appropriate localities, are not able to determine resistances 

 in absolute measure which do not vary several per cent. A stand- 

 ard of so little accuracy would not even answer the requirements 

 of technical purposes." M. Siemens, however, does not give the 

 grounds on which he bases the above assertion. 



Prof. W. Thomson, in a paper published in the ' Proceedings 

 of the Royal Society ' (vol. viii. p. 555), says, " It is impossible 

 to over-estimate the great practical value of this system of abso- 

 lute measurement carried out by Weber into every department of 

 electrical science" I have always understood that the determina- 

 tions of resistances in absolute measure by Weber's methods were 

 most accurate ; and in order to be able to answer this point more 

 definitely, I wrote to Prof. W. Thomson and asked him to give 

 me his opinion on the subject, knowing that the opinion of such 

 a distinguished physicist would have great weight — in fact, 

 would settle the question. 



Prof. Thomson's answer was the following : — 

 " There can scarcely be a doubt but that Weber's original deter- 

 mination of resistance in absolute measure (Pogg. Ann. vol. lxxxii. 

 p. 33) was considerably within one half per cent, of the truth. He 

 usedtworemarkably different methods, and obtained by means of them 

 1903 and 189'8 respectively for the absolute measure of the resist- 

 ance of one of his conductors. The details of the application of each 

 of the two methods separately present so much consistence, that the 

 possibility of so great an error as one half per cent, could not be 

 admitted in the mean result of either considered alone, unless 

 through some error in the corrections directly applied to it. Any 

 such doubt seems perfectly removed by the close agreement between 

 the two results derived from the two different methods, with different 

 instruments, very dissimilar experimental operations, and perfectly 

 distinct reductions and corrections to reduce to absolute measure. 

 The mean of the two numbers quoted above, being 19005, differs 

 by less than 014 per cent, from each of them. It is not improba- 

 ble that this mean may be within 0*1 per cent, of the truth: it is 



* Communicated by the Author. 



t Pogg. Ann. vol. cxiii. p. 91. In order to prevent mistakes I will call 

 this the 2nd paper ; the 1st paper being the original one, where M. Siemens 

 proposes mercury as a standard (Phil. Mag. Jan. 1861). 



